Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired blood disease characterized clinically by chronic hemolysis causing hemoglobinuria, bone marrow failure (BMF), and a tendency to thrombosis. PNH is characterized by clonal proliferation of a hematopoietic cell with a somatic mutation in the X-linked PIG-A gene, as a result of which glycosylphosphadityilnositol (GPI)-linked proteins are deficient on the cell surface (since one of these is the complement- protective protein CD59, this explains the hemolytic and possibly the thrombotic manifestations). In the course of the previous grant period we have determined, based on an extensive analysis of mouse models of PNH and of normal human subjects, that PIG-A mutations are necessary but not sufficient for the development of PNH. In this renewal proposal we intend to test the specific hypothesis that, once a PIG-A mutation has taken place, it does not by itself confer a growth advantage to the mutant cell; rather, expansion of the mutant clone will take place only when it gams a conditional growth advantage by virtue of a special bone marrow environment. We further hypothesize that the selective mechanism depends on the presence in PNH patients of auto reactive lymphocytes that are able to attack and damage normal hematopoietic stem cells, but not PNH stem cells. Therefore, the specific aims of this proposal are as follows. (1) To identify auto-reactive lymphocytes in patients with PNH. We plan first to establish whether these are NK cells or T cells, and then to characterize them in terms of surface markers, TCR structure and functional properties. Preliminary data suggest we are dealing with CD3+CD8+ lymphocytes. (2) To identify the target of these auto-reactive lymphocytes. The most important inferred characteristic of these cells is that they must be able to discriminate between normal and PNH hematopoietic cells. This could mean that they recognize one of the numerous GPI-linked proteins expressed on hematopoietic stem cells; or that their reactivity depends on an accessory GPI-linked protein; or that the GPI glycophospholipid anchor is it the target. These possible alternatives will be tested in turn systematically; and we have already preliminary data which tend to favor the third possibility. (3)To isolate physically these lymphocytes from PNH patients, with the ultimate purpose to develop strategies for their specific ablation. For this purpose we intend to use CD1d refolding and tetramer technology. Isolation and quantitation of the specific T cell population we are postulating would provide a new tool in the diagnosis, prognosis and follow up of patients with FNH; in addition, since we suspect that we are dealing with a small subset of very powerful auto reactive cells, it is conceivable that we may find ways to ablate them specifically, rather than using drastic immuno suppression. The significance of this work might extend to clonal disorders other than PNH.